Dual interest rates give central banks limitless fire power
With both tepid growth and inflation across the developed world following the global crisis, and a falling equilibrium rate of interest, central banks have been stuck with rates close to zero (Williams 2015). In the face of the average recession, the US Federal Reserve typically cuts interest rates by around 500 basis points (Bernanke 2016). No major central bank had the room to do so before the Covid-19 crisis hit. Economists and analysts have therefore declared major central banks to be out of monetary policy tools.
When the global economy was shut down in response to the pandemic, central banks proved that their arsenals were not in fact as empty as many economists had feared. Most analysts have focused on bond purchasing programmes as the most important monetary policy development since early March 2020 (Caballero and Simsek 2020). But they are missing the most powerful tool of all: dual interest rates. We are using the term ‘dual interest rate’ to describe a central bank policy of separately targeting the interest rate on loans and the interest rate on deposits in order to provide economic stimulus (Lonergan 2020, Greene 2020). We will argue that central banks have already introduced the tools to implement this policy, and a minor version has been introduced by the ECB in its attempt to ameliorate the effects of Covid-19 (Lane 2020).
By employing dual interest rates, central banks can go beyond targeting short-term interest rates and emergency liquidity and can provide a stimulus across the economy. They can also eliminate the effective lower bound and liquidity traps, giving monetary policy unlimited force.
The constraints of the effective lower bound
Since the global crisis, when many major central banks slashed interest rates to zero or into negative territory and grew their balance sheets with asset purchases, there has been widespread concern that central banks would lack the means to stimulate aggregate demand in the face of another crisis. With inflation stubbornly low over the past decade, most central banks were unable to hike interest rates sufficiently to rely on cutting them in the next downturn to hit their various targets. They would be constrained by the ‘effective lower bound’, the rate beyond which further rate cuts are counterproductive.
Defining the effective lower bound is more of an art than a science. One input is the equilibrium real rate of interest (r*), the real interest rate at which the economy is at full employment while inflation remains constant. Calculating r* has always been problematic, and it is questionable if the framework is even useful at very low interest rates. What we do know is that r* has declined considerably in recent decades (Yellen 2018). This means that central banks will reach the effective lower bound more frequently, and during these episodes it is likely that there will be significant shortfalls in output, employment, and inflation. A persistent inflation shortfall could erode inflation expectations, compounding the effective lower bound constraint as central banks are stuck with policy rates that are lower for even longer.
To escape the effective lower bound, central banks have considered (and employed) a number of approaches (Bernanke 2017). In the face of the Covid-19 crisis, most major central banks have engaged in long-term asset purchases. Other monetary authorities have intervened to directly target longer-term yields, such as yield curve control employed by the Bank of Japan. The Federal Reserve and the ECB have also launched monetary policy strategy reviews to consider changing their policy goals. The Fed and the ECB have also launched monetary policy strategy reviews to consider changing their policy goals. The Fed’s review resulted in the central bank adopting an average inflation target of 2%, so that shortfalls must be offset by overshooting the inflation target (Greene 2019a).
None of these approaches gives us confidence that central banks can meet their inflation targets and escape the effective lower bound. Quantitative easing has produced mixed results (Liberty Street Economics 2019, Williamson 2017). Yield curve control runs the risk of a central bank having to spend an extraordinary amount to pin yields in large and liquid markets and is difficult, if not impossible, to reverse. Targeting average inflation or price levels (or raising inflation targets) lacks credibility when most major central banks have largely failed to hit their inflation targets over the past decade, let alone overshoot them.
Some economists have argued that central banks could simply resort to even more negative nominal policy rates, but the likely existence of a ‘reversal rate’, a point at which further reductions in the policy rate result in monetary tightening, undermines this approach (Brunnermeier and Koby 2019).
The ECB’s institutional framework for dual interest rates
A far more effective approach to escaping the effective lower bound already exists and has been employed in Europe, to surprisingly little fanfare: dual interest rates. Central banks have always had multiple rates of interest. The main interest rate central banks typically aim to control is the rate at which commercial banks borrow and lend reserves to each other overnight. This rate is a benchmark for the pricing of all lending across the private sector, for discounting government bonds, and for remunerating deposits. Historically, central banks have focused on pushing the benchmark rate towards the policy rate using three main tools: supplying reserves to the market via open-market operations, setting an interest-on-reserves rate that represents what banks can earn on funds deposited at the central bank (providing a theoretical floor on rates), and using a discount rate at which a central bank will lend funds (against collateral) to banks (providing a theoretical ceiling on rates) (Keister 2012).
Over the last decade, central banks have developed a far more sophisticated array of interest rate-based tools. These have included tiered reserves as well as various ‘targeted lending schemes’ such as the ECB’s TLTRO schemes,1 the Bank of England’s Term Funding Scheme (TFS), and the Bank of Japan’s targeted lending programme introduced following the Fukushima emergency (Lonergan 2019a). The combination of targeted lending at interest rates below the IOR and the use of tiered reserves imply that central banks can now independently target lending rates and deposit rates. In theory, conventional monetary policy leaves the net interest income of the private sector unchanged. The effects of changes in the policy rate work through inter-temporal substitution of consumption (a price effect) or through the differential marginal propensities to consume for borrowers and lenders. By contrast, the effects of dual interest rates are unambiguously positive because the central bank can simultaneously raise the income of both borrowers and lenders. This is the equivalent of monetary rocket fuel (ECB 2020a).
So far, however, there is only one example where this approach has actually been employed. In tweaking its Targeted Longer-Term Refinancing Operation (TLTRO) programme in March 2020, the ECB adopted dual interest rates. The ECB began explicitly and separately targeting the interest rate on lending and the interest rate on deposits. As a result, the ECB funds banks at one interest rate (contingent on net new lending conditions being met) which is independent of the interest on reserves (IOR) rate. Banks can borrow under this scheme at interest rates as low as -100 basis points. Simultaneously, the interest rate on reserves (or deposits held by commercial banks at the ECB) is determined independently under the ECB’s tiered reserve system. The ECB’s ‘dual reserve system’ remunerates a portion of the commercial banks’ excess reserves at zero, and the remaining balance at the ECB’s deposit rate (currently set at -50 basis points) (ECB 2020b).
To be clear, the ECB is targeting the rate at which banks can fund lending independently of the rate at which reserves are remunerated. By continuing to cut the interest rate through the TLTRO programme, while leaving the average IOR rate unchanged, the interest rate on loans in the euro area should fall without a commensurate decline in interest income on deposits. This radical departure from monetary convention has not been lost on the ECB, as the chief economist, Philip Lane has made explicit:
“Banks can borrow at the most favourable rates we have ever offered, provided that they continue to do their job of extending credit to the private sector. An important innovation is that, by setting the minimum borrowing rate at 25 basis points below the average interest rate on the deposit facility, we are effectively lowering the funding costs in the economy without a generalised reduction in the main traditional policy rates” [italics added] (Lane 2020)
This policy could be taken much further. For example, the ECB could cut the interest rate on transfers within the TLTRO programme to -400 basis points and could extend the duration of loans to 18 months or longer. Furthermore, instead of linking preferential interest rates to net new lending (particularly when private sector balance sheets are already stretched), the ECB could link an aggressive reduction in the programme rate to the re-pricing of existing loan books. This would create a stimulus for borrowers across the economy.
Normally, a central bank has to choose whether to benefit borrowers or savers when it either hikes or cuts the main policy rate. But with dual interest rates, the central bank can provide a stimulus to borrowers while also extending one to savers at the same time. In addition to cutting the lending rate deep into negative territory, the ECB could raise the average interest rate banks receive on reserves. Tiered reserves are particularly relevant here. Under a single rate regime, raising the IOR rate raises lending rates across the economy. But under tiering (such as that which the ECB has adopted), the zero rate on a portion of excess reserves could be raised to say 2% or 3%, conditional on banks passing a share of this higher interest income on to customers.
There is no lower bound
What becomes clear is that under a dual interest rate system, monetary stimulus has no practical limit (Greene 2019b). In addition to providing a stimulus across the entire economy, dual interest rates remove the effective lower bound. Conventional negative nominal interest rates provide a stimulus for borrowers only. Negative nominal policy rates pejoratively impact depositors, pose profound challenges for financial intermediation, and run the risk in extremis of creating an incentive for the private sector to hold physical cash.
These constraints disappear once dual interest rates are employed. There is no lower bound to how far central banks can reduce the interest rate at which they can lend to banks, as the ECB has now demonstrated. By making this lending ‘targeted’ (i.e. subject to eligibility criteria), the scope for free-riding is limited if not impossible (even more so if funding terms are linked to a repricing of existing loan books, as we suggest).
The legitimate concerns about the effects of low (or negative) interest rates on depositors (highlighted in debates about the impact of monetary policy on inequality, and raised by the German Constitutional Court) are also addressed by dual interest rates. Central banks can use tiered reserves to support rates of interest on deposits. In effect, central banks can use dual interest rates to explicitly target banking sector net interest margins in the application of policies.
The monetary policy implications of these tools cannot be overstated. Where dual interest rates are employed, there can be no liquidity trap, and no reversal rate. The liquidity trap only makes sense in a world where base money creation occurs through the acquisition of an asset by the central bank, leaving the net wealth of the private sector unchanged. Under dual interest rates, the central bank is, in effect, making a money transfer to the private sector, which raises net wealth. Intriguingly, this argument was made very succinctly by Milton Friedman in his famous 1968 AER address as a reason why monetary policy can always be effective even at very low interest rates:
“This revival [in a belief in the potency of monetary policy] was strongly fostered among economists by the theoretical developments initiated by Haberler but named for Pigou that pointed out a channel – namely, changes in wealth – whereby changes in the real quantity of money can affect aggregate demand even if they do not alter interest rates. These theoretical developments did not undermine Keynes’ argument against the potency of orthodox monetary measures when liquidity preference is absolute since under such circumstances the usual monetary operations involve simply substituting money for other assets without changing total wealth. But they did show how changes in the quantity of money produced in other ways could affect total spending even under such circumstances” [italics added] (Friedman 1968).
Friedman is not explicit about what he means by producing money “in other ways”. It is likely he had in mind a simple cash transfer. Relative to asset purchases, dual interest rates involve a relatively modest transfer of net income to the private sector, the effects of which can be amplified by bank lending.
Are there constraints on dual interest rates?
At first glance, dual interest rates seem too good to be true. Economists are taught that there is always a constraint. In the case of dual interest rates, the constraint is unclear. A combination of negative interest rates on targeted lending and positive interest rates on tiered reserves can deliver any level of nominal demand required to raise the inflation rates. Monetary policy is given new breath when inflation is ‘too low’. The same applies when inflation is ‘too high’; the interest on tiered reserves can be set to shrink reserves in the same way as it can be set in order to expand them, and borrowing rates can be raised as well. The challenge for central banks becomes devising an optimal calibration. What is the r* for the TLTRO programme? What average interest rate should be applied to tiered reserves?
Central bank ‘equity’ is also not a constraint. One significant difference between a policy of dual interest rates and the more ‘plain vanilla’ asset purchases by central banks (QE) is the effect on a central bank’s balance sheet. When central banks engage in asset purchases there is usually no accounting change to their equity (because they have created an accounting liability, bank reserves, and have a corresponding asset). Under dual interest rates, new operations will create negative net interest income for the central bank. It is of course true that central banks asset purchase programmes expose them to mark-to-market losses, so the distinction is somewhat semantic. Chris Sims has made a similar point, albeit to express concern over the size of central bank balance sheet and the political consequences of ‘losses’ (Sims 2016).
Even in accounting terms, central banks may have far greater equity than appears at first glance. The accounting treatment of bank reserves as ‘liabilities’ resides in the mapping of commercial bank accounting to central banks and makes little sense. No other institution can create electronic money, so we do not have accounting conventions to deal with base money. Central bank reserves are not financial liabilities in any meaningful sense. They can be created at will, and at any ‘price’ the central bank wants, and ‘default’ has no real meaning in this context.
As economists, we have typically focused on the asset side of a central bank balance sheet from the perspective of the ability to implement monetary policy through open-market operations. Even mark-to-market losses are relevant if they hinder a central bank’s ability to reverse the expansion of the monetary base. Similarly, under a single IOR regime, it is possible in the future that central banks’ net income is significantly negative. Tiered reserves limit this concern. If central banks in the future have reserves in excess of assets and need to shrink the stock of reserves while raising interest rates, an ability to charge negative interest rates on required reserves will be valuable.
The politicisation of monetary policy
A number of economists have talked about the blurred lines between fiscal and monetary policy (Bean 2017, De Groen et al. 2016). This is relevant to both the politicisation of monetary policy and its perceived legitimacy and independence. In Europe, the issue is even more acute as legality is potentially an obstacle to any policy innovation, as highlighted by the recent ruling by Germany’s constitutional court.
Dual rates meet both abstract and theoretical definitions of monetary policy, but also mirror more closely the traditional tools of central banks, which may appease those concerned by the perceived political implications of more radical prescriptions. Monetary policy is defined most clearly in the abstract (as changes to the quantity and price of base money). In institutional terms, the broadest definition of monetary policy is the set of policies legally implemented by the central bank. To the extent that a significant number of the major central banks have already put in place the mechanisms by which to implement dual interest rates, it falls, rather uncontroversially, into the defined realm of monetary policy.
Many commentators seem to object to central banks moving away from setting interest rates. To that extent, faced with the option of an array of unconventional policies which we may be required to draw upon post-Covid-19 (such as various forms of helicopter money) (Lonergan 2019b), dual interest rates may prove both highly effective and the least unsettling politically.
Employing dual interest rates, whereby central banks independently target bank funding rates and the interest rate of reserves with the objective of influencing deposit and lending rates independently, is a major innovation in monetary policy. Dual interest rates go beyond just targeting short-term interest rates or providing emergency liquidity to deliver an outright stimulus. This tool eliminates the effective lower bound and the risk of a liquidity trap, giving monetary policy new life in a world in which the real equilibrium interest rate is already low, is trending lower, and aggregate demand is likely to remain depressed in the wake of a pandemic and lockdown measures.
Importantly, central banks across the world already have the infrastructure in place to implement a policy of dual rates and this approach has already been used by the ECB, albeit timidly. Many developed economy central banks have tiered reserves, which allows the central bank to protect the commercial banking sector’s margins when interest rates are very low. This could also be used to encourage higher deposit rates for the private sector more broadly, without undermining the setting of interbank rates.
Some might argue that employing dual interest rates falls outside a central bank’s remit and blurs the line between monetary and fiscal policy too much. But a policy of dual interest rates falls clearly under the remit of central banks and monetary policy. It involves the creation of bank reserves (the defining feature of monetary policy) and the price of those reserves; interest rates.
Nominal demand will prove stubbornly depressed in the post-Covid-19 and post-lockdown world, and there are already signs that fiscal authorities are beginning to drag their feet, as evidenced by the likely expiry of wage guarantees in the UK, and in the US Congress’s difficulty agreeing a fourth stimulus package. As is so often the case following a crisis, we expect governments to face increasing calls for fiscal restraint in the face of unprecedented budget deficits. With dual interest rates at their fingertips, central banks will have no excuse for staying on the sidelines and should step in with aggressive, overwhelming support.
Bean, C (2017), “Central banking after the Great Recession”, Wincott Lecture, Chatham House. 28 November.