Central bank losses and monetary policy In Theory

Central bank losses and monetary policy In Theory and Practice Tomáš Holub Monetary Economics Seminar 9 November 2015

Motivation – the Czech case (i) • The CNB was in negative equity from 1998 until very recently. • The negative trend has reversed since 2008. 2

Motivation – the Czech case (ii) CNB‘s balance sheet structure • CNB‘s assets are dominated by FX reserves, which are not matched by FX liabilities. This implies a high exposure to the exchange rate risk (and in some stages also to losses related to the catching-up process). • Excess liquidity is being sterilised via 2 W reverse repo 3

Motivation – the Czech case (iii) • The losses arose mainly due to FX revaluations. • Trend real exchange rate appreciation in a low inflation environment → nominal exchange rate appreciation before the crisis. 4

Motivation – a few other examples of negative equity Equity / total assets • The Czech Republic is not the only country that has experienced negative central bank equity. • Other well known examples are Israel, Chile and Slovakia… 5

Motivation – global country sample • There are CBs that have experienced losses or even negative equity, even though they are not very numerous. • Historically, one can find both harsh and benign cases in terms of policy outcomes (see below). 6

Outline • Does it matter fiscally? • Does it matter theoretically for the independent policy conduct? – Financial constraints (central bank vs. commercial banks) – Political-economy constraints • Does it matter in practice? – Do policy-makers care? – Selected harsh cases – Selected benign cases – The available empirical work • Conclusions and implications for the institutional design 7

Does it matter fiscally? Primary deficit Interest on debt Profit transfer from CB Government debt issuance • Profit transfers from the central bank limit the govt‘s debt issuance needs; and thus yes, CB finances do matter fiscally (see the lecture). • Under sensible profit distribution rules, central bank losses should reduce current and/or future profit transfers to the government. • If a central bank with negative equity gets recapitalised by the gov‘t, it is like a negative profit transfer, increasing the debt issuance needs. • Example 1: The CNB has paid no dividends to the gov‘t since 1990. 8

Does it matter for independent policy conduct? • …thus, the society does have a legitimate interest in the financial performance of the central bank. • But central bank losses are not necessarily a result of bad policies; they may often be a by-product of good policies that pay off for the society in the long-run. • Moreover, the experience has taught us that it is desirable to delegate monetary policy-making and certain other functions to independent institutions isolated from the election cycle. • The question here is thus whether central bank‘s financial weakness affects significantly its ability to conduct policies independently. 9

Does it matter financially? Stylised Balance Sheet of a Central Bank Assets Liabilities Net foreign exchange assets (NFXA) Currency (CU) Gov‘t bonds held by CB (GB) Sterilisation of liquidity (SL) Lending to banks (LB) Equity (EQ) • Central banks differ significantly from commercial banks and other private enterprises. • They are typically not subject to the standard bankruptcy procedures and may thus operate even with negative equity. • Central banks can also hardly become illiquid, because they are the ultimate issuer of the domestic currency, and they can thus settle their obligations by increasing their liabilities – currency or bank reserves. • CBFS extends well beyond their accounting equity, because the issued currency behaves in many respect 10

Does it matter financially? Stella (2010): ”In a floating exchange rate environment, currency and non-interest bearing deposits at the central bank play a financial role very similar to capital. “ Bini Smaghi (2011): ”The privilege to issue legal tender gives central banks an additional financial buffer in the form of seigniorage income. … The seigniorage income expected for the future constitutes an implicit financial buffer that needs to be considered when assessing the economic capital of a central bank. In this regard, the net present value of such future income can be seen as a sort of ‘franchise capital’, as it reflects the value of the ‘franchise’ to issue legal 11 tender. “

Does it matter financially? Stylised P&L of a Central Bank Incomes on Net FX assets = yfx*NFXA on Gov‘t bonds = ygb*GB on Lending to banks = ilb*LB Profit π = Incomes - Costs Monetary income +return on equity Cost of carrying FX reserves (cost of sterilising it + revaluation losses) Costs on Currency = 0 Sterilisation cost = i*SL on Equity = 0 Operating outlays (OL) Valuation losses (or gains) on GBs + possible preferential lending to gov‘t ( ) Losses on lending to banks (e. g. due to solvency support) • Central banks are structurally profit-making … unless their financial results get undermined by some unpleasant yield differentials (+operating outlays). 12

Does it matter financially? „Structural“ profitability The financial „tipping“ point • A structurally profit-making central bank can get out of negative equity without recapitalisation, unless it gets beyond some tipping point. 13

Formalized analysis of CBs‘ financial constraints • Sims (2004): focuses on “fiscal aspects of CB independence“ • Bindseil, et al. (2004): explore the role of CB capital in ensuring the focus on price stability, develop a simple model of the relationship between a CB‘s balance sheet and its inflation performance • Ize (2005): derives a concept of “core capital” (a function of the central bank’s operating expenditures and the carrying cost of its international reserves) as the minimum capital needed by a central bank to ensure the credibility of its inflation target • Buiter (2006): concepts of the “financeability” of inflation targets, and of their “independent financeability” by central banks 14

Does it matter for other (i. e. ”softer“) reasons? • Fry (1993): quasi-fiscal activities undermine the independence of CBs and ultimately also their MP objectives, as the resulting losses must eventually be covered by an expansion of CB money • Stella (1997, 2005): large negative net worth of a central bank is likely to interfere with price stability • Cargil (2005): Bo. J had put too much weight on its financial results, an undesirable policy constraint • Stella and Lönnberg (2008): the term “policy insolvency” as opposed to “technical insolvency” • Cukierman (2011): CB capital important for preserving its policy independence, although other institutional aspects matter, too 15 -----------

The direction of causality (? ) CB finances 1 • We focus on link 1 – can weak finances force a CB to inflate? • Much less obvious and definitely much less explored – relies mainly on ‘soft’ arguments (reputation, politicaleconomy). • But some empirical papers claim that the link is there. • Potentially strong nonlinearity (threshold-effects; 2 Inflation CB • Link 2 is obvious and well explored in the literature, going back e. g. to Cagan (1956) – see the lecture. • Non-linearity relevant mainly for high inflation rates. ? 16

Does it matter for other (i. e. ”softer“) reasons? • Can weak finances force a CB to inflate in order to move up the seigniorage ”Laffer curve“ (or to pursue some other policy goals less vigorously, e. g. in the financial stability domain)? • Political pressure: the government may try to replace a loss-making central banker, possibly as a scape-goat, or to limit CB independence. • Need of recapitalisation: if a CB is forced to ask the government for recapitalisation, it may get into a quid-proquo game. • Avoiding the above: fearing the adverse political consequences, a CB may itself decide to avoid potentially loss-making activities, to have less ambitious inflation targets or impose high unremunerated reserve requirements. 17

Do policymakers care? • Fukui (2003), Bank of Japan: discussion of the 1 st QE episode • de Paula Gutierrez (2005), Central Bank of Costa Rica: “a negative net worth…remains our greatest challenge“ • Shirakawa (2010), Bo. J: recent QE episode; stresses that preference was in the end given to policy goals • Bini-Smaghi (2011), ECB: capital position of a CB, its profitability and the degree of financial risk protection seem to be crucial elements that contribute to its credibility, hence facilitating monetary policy (but also recall: ”the privilege to issue legal tender gives CB an additional financial buffer in the form of seigniorage income“) • Fischer (2011), Bank of Israel: costs of FX reserves accumulation 18

Example 1: ECB‘s Convergence Reports ECB Convergence Report 2010: ”In particular, any situation should be avoided whereby for a prolonged period of time an NCB’s net equity is below the level of its statutory capital or is even negative, including where losses beyond the level of capital and the reserves are carried over. Any such situation may negatively impact on the NCB’s ability to perform its ESCB-related tasks but also its national tasks. Moreover, such a situation may affect the credibility of the Eurosystem’s monetary policy. Therefore, the event of an NCB’s net equity becoming less than its statutory capital or even negative would require that the respective Member State provides the NCB with an appropriate amount of capital at least up to the level of the statutory capital within a reasonable period of time so as to comply with the principle of 19 financial independence. “

Example 2: RZBN capital injection • In 2004 the Reserve Bank of New Zealand (RBNZ) got a new mandate in FX policy that gave it greater freedom to intervene for exchange rate management reasons. As the RBNZ was aware of the potential financial risks associated with this new mandate, it approached the government with a request for a capital increase. • ”If the Bank is undertaking intervention under Section 16 of the Act, then it must be able to absorb any short-term marked-to-market losses from its intervention strategy on its balance sheet, without appearing to jeopardise its solvency. It is thus the expected size of these (likely temporary) marked-to-market loss situations which best determines the level of capital needed to be invested in the business. . then a capital injection of around NZD 1 billion is prudent. ” http: //www. rbnz. govt. nz/finmarkets/foreignreserves/intervention/014720

Example 3: recent RBA capital injection • The Australian government has given the Reserve Bank of Australia (RBA) A$8. 8 billion ($8. 5 billion) to enhance the RBA's "capacity to conduct its monetary policy and foreign exchange operations", blaming the strength of the Australian dollar and "the determination of the previous government to take extraordinary dividends from the Reserve Bank" for the depletion of its reserves. • The cash transfer will bring the Reserve Bank Reserve Fund (RBRF) up from 3. 8% of the RBA's assets at risk, to 15% – and has been made at the request of RBA governor Glen Stevens, according to Australian treasurer Joe Hockey. • Hockey claimed in his statement today that the government's A$500 million dividend in 2012– 13 was taken "against the advice of the Reserve 21 Bank…Restoring the RBRF to 15% of assets at risk will,

Selected harsh cases • National Bank of Liberia (NBL) in the 1980 s and 1990 s: this case demonstrates that circumstances can go so awry that central bank liabilities may cease to be money and the central bank can default on obligations in its own currency. The NBL was expelled from the local cheque clearing house for repeated settlement failure. • The Philippines: a weak balance sheet coincided with high inflation in the years prior to the 1994 reestablishment of the central bank with a clean balance sheet. • The Central Reserve Bank of Peru (CRBP): had a long history of mainly quasi-fiscal losses until the introduction of a new law in 1993, financed primarily by money creation, thus leading to a hyperinflation. • Central Bank of Costa Rica (CBCR): quasi-fiscal operations and revaluation losses on the external debt 22

Selected benign cases • Typically catching-up economies, losses due to FX revaluation (with some past quasi-fiscal losses), inflation targeting regime, manageable fiscal situation, medium-to-high degree of CB independence and transparency. 23

Harsh versus benign cases Three conditions for a benign outcome: 1. Policy frameworks need to be sound… 2. The reason for losses is politically acceptable… 3. Long-term financial strength must be undoubted. 24

Empirical work – literature review • Stella (2003): Mean inflation for the ’weak‘ CBs was 26%, twice as high as for the ’strong‘ group, this difference being significant at all standard levels. See also Stella (2008; 2011) • Ize (2006): Average inflation rate was 9. 5% for the ’weak‘ central banks and 3. 5% for the ’strong‘ ones • Klüh and Stella (2008): Found a relatively stable and robust negative relationship between CBFS and inflation. But only a relatively strong BS impairment would result in significantly higher inflation • Adler, et al. (2012): Found out that CBFS can be a statistically significant factor explaining large negative interest rate deviations from an estimated forward-looking Taylor rule. • Perera, et al. (2013): Claim that there is a statistically 25

Stylized facts – correlations correlation =-0. 04 correlation =+0. 04 • There is no apparent correlation between the standard accounting measures and inflation. • For Klüh and Stella‘s (2008) broader CBFS 1 measure, the correlation 26 is -0. 25.

Benecká, et al. (2012): estimation results summary • Some estimates show a statistically significant negative relationship between CBFS and inflation, but this link appears rather weak and not as robust as suggested by the other papers. • There is no CBFS measure that would consistently perform best. • GMM, which is probably the most appropriate method, shows significance of ROAA, i. e. a rather short-term 27

Benecká, et. al: non-linearity analysis (OLS estimates) … with sampling of countries by inflation … with sampling of countries by CBFS • Significant outcomes often found only with the weakest 28

Conclusions and implications for the institutional design • CBFS is a relevant topic, although policy should receive primacy. • CB losses usually do have fiscal implications. • They may adversely affect independent policy conduct, if a financial tipping-point is reached, or due to adverse political dynamics. • The harsh historical examples often reflect (quasi-)fiscal losses in countries with major institutional weaknesses. • Strong policy frameworks, losses associated with policy success (e. g. low inflation and currency appreciation), and undoubted long-term financial strength underlie the benign cases. 29