Monetary Policy: Inflation Targeting 2.0

1. The Future of Monetary Policy: Inflation Targeting 2.0

The problem: Central banks can't ignore asset prices... Monetary policymakers' attitude of benign neglect towards asset price developments has not worked. The combination of runaway house prices and household and financial sector leverage in the US and elsewhere resulted in the worst slump in living memory. The orthodoxy in academia and central banks was that monetary policy could simply ‘mop up' after a bubble has burst, but the slump proved too severe for conventional monetary policy to manage. Central banks had to use unconventional tools such as massive balance sheet expansion to stem the tide, risking high inflation down the road (see "The Big Easy", The Global Monetary Analyst, July 28, 2010).

...but ‘leaning against the wind' may compromise the inflation target: We know from experience that low inflation does not ensure financial stability. The financial crisis and the Great Recession followed a period of low and stable inflation; as did the Great Depression in the 1930s and the bursting of the Japanese bubble in the 1990s. With the benefit of hindsight, the optimal response for central banks prior to the recent crisis would have been to lean against the wind sufficiently to choke off the build-up of financial imbalances (see "Inflation Targeting: The Trial", The Global Monetary Analyst, July 21, 2010). However, this would probably have come at the expense of the inflation goal since inflation was largely on target. That is, occasionally there may be a short run trade-off between the inflation objective and the maintenance of financial stability.

The solution? Add a tool: Note that the trade-off arises because the single tool - the interest rate - is employed to pursue financial stability in addition to inflation. The level of the policy rate consistent with achieving the short-term inflation target may be different from the interest rate level that would achieve reining in asset price misalignments. Hence, it may be possible to cut the Gordian knot by adding a tool with which to pursue financial stability, while the traditional interest rate tool remains devoted to the pursuit of the inflation objective. A growing consensus in central banks, international organisations and academia posits that such a tool exists. In the words of the Bank of England's Charles Bean: "[W]ith an additional objective of managing credit growth and asset prices in order to avoid financial instability, one really wants another instrument that acts more directly on the source of the problem. That is what ‘macro-prudential policy' is supposed to achieve".

Prudential policy goes macro... To the extent that the behaviour of financial institutions is instrumental in the build-up of asset price bubbles, it makes sense to consider regulatory policy as an antidote to financial imbalances. In such a case, the instruments will also be regulatory in nature, thus focusing on capital and liquidity. However, the ‘macro' dimension is crucial. That is because micro-level regulation involves a ‘fallacy of composition' in times of crisis. A microprudential regulator would be satisfied with an undercapitalised institution restoring capital adequacy by either raising capital or selling assets. But when the financial system is in distress, a simultaneous attempt by institutions to sell their assets would make the situation worse. According to Fed Chairman Bernanke, "The Federal Reserve is working...to move from an institution-by-institution supervisory approach to one that is attentive to the stability of the financial system as a whole".

...and dynamic: Why ‘macroprudential' must be ‘countercyclical': Macroprudential regulation should be dynamic for very similar reasons. Micro-level regulatory policy is procyclical as adherence to a fixed capital adequacy ratio could trigger asset sales as described above. Rather than reinforcing the inherent procyclicality of the financial system, regulation focused on system-wide stability should counteract it. How? Regulatory tools should be varied in a countercyclical fashion - the authorities should increase capital and liquidity requirements in an upswing and lower them in a downturn.

The new tool could improve a fundamental policy trade-off: There is another perspective on macroprudential policy. A central bank generally faces a trade-off between stabilising inflation (minimising the volatility of inflation) and stabilising output (minimising the volatility of output). The more it seeks to maintain inflation at a certain level, the more activist the use of the interest rate will have to be, and the more volatile output is going to be - and vice versa. Now, asset price misalignments - due to the familiar boom-bust dynamics - increase both output and inflation volatility. Hence, mitigating the build-up of asset price misalignments should shift the output volatility-inflation volatility trade-off inwards. In principle, the new policy tool therefore holds considerable promise: it may be able to improve a fundamental trade-off in economic policy.

How new is the ‘new' policy regime? Notwithstanding the introduction of a new tool, the overall monetary policy framework should not change drastically. IT as it is currently understood and practiced - maintaining output growth around a level consistent with low and stable inflation - will remain the main mission of monetary policy. The hoped-for stabilisation of the financial system through the use of a macroprudential tool would help the CB in this mission, not change the mission itself. In short, evolution rather than revolution is the name of the game: Inflation Targeting 2.0 is IT plus the macroprudential tool, not a fundamentally different monetary policy strategy.

2. The New Tool: A Sneak Preview

A new tool requires a certain amount of groundwork. The following issues will need to be resolved. 1. Objectives: what is the new tool trying to achieve? 2. Instruments: defining the new tool. 3. Indicators: what to use as a compass for the tool. 4. High-level principles: how to use the tool.

Objectives: Introduction of a new tool should, of course, be accompanied by a definition of clear objectives in pursuit of which the new tool should be used. Should macroprudential policy attempt to consign asset price bubbles to history? This would probably be unnecessarily ambitious, and risks too many negative side-effects. Rather, in the words of Paul Tucker of the Bank of England, the macroprudential tool should aim to strengthen the "dynamic resilience of the banking system". This is a more modest aim, but also a more meaningful one.

Instruments: Once we've decided what we want to achieve, we can think about which tool to employ. Solvency (ultimately, having sufficient capital that would absorb losses) and liquidity (a mismatch in the maturity of assets and liabilities) are natural starting points when thinking about the ‘resilience' of the banking system. In what follows, we will focus on capital as the main instrument; we think that capital requirements are likely to have more wide-ranging macro implications than liquidity requirements.

Indicators: Research at central banks, the BIS, the IMF and elsewhere, mainly focused on industrialised economies, suggests that measures of private credit growth, and to a lesser extent monetary aggregates, are good indicators of asset price booms. While the discussion in the central banking community is far from settled, at the moment it seems that credit, asset growth or leverage variables have the edge over monetary aggregates.

Principles: Rules yet again: Given enough time, the macroprudential tool will likely take the shape of a macroprudential rule (MR), similar in spirit to the familiar Taylor Rule currently used as a yardstick by monetary authorities for interest rate setting. Again as with interest rate setting, macroprudential policy will be exercised with a considerable amount of discretion. That is, the macroprudential rule will not be mechanically adhered to, as incomplete knowledge of the evolving structure of the economy, of the impact of the new tool on the economy, as well as unexpected shocks will require policymakers to exercise judgment when applying the new tool.

Where the rubber meets the road: The new tool will most likely take the form of a rule whereby financial institutions' capital adequacy ratio will be varied countercyclically according to some measure of the banking sector's aggregate balance sheet. While reasonable enough, this is not sufficiently specific for practical implementation. Most importantly, banks are not all alike; differences in size and lending book growth will, at any given point in time, make for different contributions to aggregate lending growth. So, how would the rule work for an individual bank?

•           Establish the macrofinancial indicator(s) that best reflects systemic financial risk, e.g., credit growth, the growth of bank assets, or a measure of leverage.

•           Establish (forecast) the evolution of this indicator that would be consistent with the inflation target (say a level of X*), including a threshold ‘d' around it to reflect uncertainty.

•           If a bank exceeds the upper bound of the indicator (X* + d), its capital requirement increases above the Basel - mandated microprudential requirement, k0. That is, as soon as the upper bound is exceeded, the capital requirement jumps from the microprudential minimum to a higher level and from then on increases linearly with asset growth.

•           To provide a numerical example, suppose the microprudential capital requirement is 8%, ‘target' bank asset growth is 10%, the ‘allowance' is 2% and capital requirements increase by 0.5pp for each 1% excess growth in bank assets. Then a bank with 16% asset growth will need to set aside 10% of assets (= 8% minimum requirement plus 4% excess asset growth times 0.5).

3. Inflation Targeting 2.0: Interaction of Macroprudential and Interest Rate Policy

This is, of course, uncharted territory. Reassuringly, most of the time, MR and interest rate rule would move in the same direction. In an expansion, interest rate policy will be restrictive (interest rates above ‘neutral') and capital requirements raised (Quadrant 1 in Exhibit 5 of the note). In a downturn, the objective is to cushion the economy; hence, interest rates would be low (below ‘neutral') and capital requirements would be lowered (Quadrant 3). This is how, in theory, Inflation Targeting 2.0 is supposed to reduce the amplitude of cyclical fluctuations in the economy and asset markets.

The really interesting cases arise in ‘non-standard' situations (Quadrants 2 and 4). For example, a supply shock - an abrupt, substantial increase in the price of an input to production - could be countered by a relaxation in capital requirements and an increase in the policy rate (Quadrant 2). The former would lower the cost of finance, another important input in production, while the latter would restrain nominal spending and thus help to prevent ‘second-round effects'. This might have been the right policy mix to pursue in the early stages of the crisis in 2007-08, when commodity prices were galloping but the financial system was fragile, had the macroprudential tool been available then. In an alternative example, the central bank may want to stimulate demand but remain concerned about household leverage. It could then keep interest rates low to encourage spending but maintain tight credit conditions to discourage indebtedness - a situation we've called ‘Spend but don't borrow' (Quadrant 4).

4. Brave New World? Life Under Inflation Targeting 2.0

4.1. The Economy and Asset Markets

What does the adoption of this new tool imply for monetary policy, the economy, interest rates and other asset prices? As yet, we know very little about the details of the macroprudential instrument and hence how Inflation Targeting 2.0 will look (let alone when it will be implemented). The following thoughts are therefore conjectures, rather than forecasts. And of course, the ‘all else equal' proviso applies.

•           Increased macroeconomic stability: If it works as intended, the macroprudential tool should facilitate hitting the inflation target over a long horizon, by preventing the large undershoots that come with an asset price bust - induced recessions. That is, output and inflation should be more stable over the long run. In this crucial sense, Inflation Targeting 2.0 should be an improvement over the current regime, which (may have) promoted stability in the short run but was unable to prevent the build-up of large imbalances that ultimately proved very costly.

•           Less activist central bank interest rate policy: In general, interest rate and macroprudential policy should reinforce each other. This is because: i) the effectiveness of the transmission mechanism will be enhanced as the financial system becomes more resilient; ii) there will be less need to cut interest rates drastically if the prevention of financial imbalances means fewer busts; and iii) central banks that already lean against the wind will need to do less upon the introduction of the macroprudential tool. In short, if all goes according to plan, introduction of the macroprudential tool should ‘lighten the burden' of monetary policy, hence requiring less use of the interest rate tool.

•           Lower volatility of short-term interest rates: Less activist (and hence more predictable) interest rate policy as well as more stable interbank markets should lower the volatility of front-end rates.

•           Lower volatility of long-term interest rates: Better anchored output and inflation over a long horizon should stabilise long bond yields (and lower term premia).

•           Less cyclical credit spreads: Credit spreads are tight in an expansion and wide in a downturn. As discussed, macroprudential policy will likely arrest the (growth in the) supply of credit in a boom and therefore increase the price of credit, and vice versa in a recession. This means that in an expansion (Quadrant 1) the macroprudential tool should imply wider credit spreads, while in a recession (Quadrant 3) credit spreads should be tighter.

•           Less cyclical return on bank equity: The new tool will probably constrain the expansion of credit in the upswing and cushion its fall in the downturn. This should make for less cyclical bank profits and hence less cyclical bank equity returns.

4.2. Problems and Challenges

"The best-laid plans of mice and men often go awry". Here are some of our concerns:

Will markets go along with it? The macroprudential tool should be varied countercyclically. Yet policymaker-mandated capital requirements may not always be a binding constraint. In particular, we are sceptical whether markets will be comfortable with financial institutions reducing capital ratios in a downturn.

Will it be enough? Adding a macroprudential rule may not obviate the need for the central bank to lean against the wind. Fed Chairman Bernanke warns that "if...reforms...prove insufficient to prevent dangerous build-ups of financial risks, we must remain open to using monetary policy as a supplementary tool for addressing those risks". There is a risk that central banks find themselves in the worst of all worlds: having a new tool with all the risks this poses while still having to employ the traditional interest rate tool against asset prices - possibly at the expense of the inflation target and the overall transparency of monetary policy.

Disappointed expectations and central bank credibility: A related problem: although financial stability cannot be defined as easily as price stability (see our discussion above), the absence of it is always obvious. The new tool essentially raises the stakes for central banks: politicians and the public will be much less forgiving of monetary authorities when the next crisis occurs. If Inflation Targeting 2.0 were to fail in preventing financial instability, the damage to central bank credibility could be irreparable.

Informational requirements: It is not implausible to suggest that varying capital requirements as a function of aggregate measures alone may be ineffective. Banks may respond by restricting lending to low-profitability, low ‘exuberance' sectors and increase lending to high-profitability, exuberant sectors, with no overall reduction in the riskiness of the financial sector. The macroprudential instrument may then have to become more granular, differentiating exposure by sector, e.g., mortgages versus corporates, etc. (‘risk weights'). It is clear that this requires the authorities to collect and process enormous amounts of information - which then need to be assessed correctly in order to subsequently become the basis for policy action.

Detail versus transparency: Further, the higher the degree of granularity, the lower the transparency.

Mission creep: In the same context, there is a danger that policymakers will succumb to the temptation to micromanage credit and financial conditions - or be perceived as doing so.

Balance: As with every policy tool, the authorities need to take care to strike the right balance. In this case, it is between the extremes of smothering the economy by making too frequent/stringent use of the policy tool and the risk of doing too little. (The same is true for the degree of granularity at which the policy tool should be applied.)

Uncertainty about the stance of monetary policy: Much of the effectiveness of monetary policy comes from the public understanding the central bank's intentions and current stance. In some cases, the two policy tools may move in opposite directions in ways that are difficult to communicate. This could decrease the transparency of the policy regime and thus its effectiveness. Ultimately, of course, the question is whether the costs of occasional lack of clarity on behalf of policy are outweighed by the benefits of having an additional policy tool.

The danger of policy mistakes remains... Perhaps the most important issue when it comes to the application of MR is identification of the drivers underlying changes in financial conditions (e.g., increases in leverage or credit growth). For example, a pick-up in house price and credit growth may occur because of increases in productivity rather than ‘irrational exuberance'. In such a case, tightening financial conditions would be the wrong response. This gets us back to the rules versus discretion debate. It becomes clear that given the enormous uncertainty surrounding the identification of the drivers of financial conditions, following rules rigidly is inadvisable, and discretion is required.

...as does the danger of complacency and overreliance: Prior to the crisis, economists, central bankers and politicians were congratulating themselves for having defeated the business cycle. Having one more tool at their disposal might make central banks feel more powerful than they actually are - in fighting yesterday's battle. The next crisis is sure to come, and it will be just as unexpected as the previous one

Surprising CPI Increase

CPI surprisingly increased by 3.6%Y in September, much higher than 2.6%Y in August.  It was also way higher than our forecast of 2.8%Y and the consensus forecast of 2.9%Y. On a sequential basis, CPI rose by 1.1%M, compared to 0.3%M in August. Food prices were the main driver of the surging CPI in the month, increasing 13.0%Y in September (versus 5.7%Y in August), or 6.6%M (versus 1.7%M in August). The Chuseok holiday in the month could be one of the seasonal factors pushing food prices higher. Inflation in the prices of other items remained relatively contained. Core CPI (excluding food and energy) increased by 1.9%Y (versus 1.8%Y in August) or 0.3%M (versus 0.2%M in August). 

Food Prices Were the Driver

Food prices rose by 13.0%Y in the month, the fastest level in almost six years.  Food items (including eating out) accounted for 27.0% of the total CPI basket, and this was the biggest driver of CPI in the month. The rise in food prices in recent months was due mainly to the supply shock of vegetables, fruit and fishery items caused by the abnormal weather in the summer season. The increase in demand during the Chuseok holidays also pushed prices higher. Apart from food, the prices of other categories remained relatively contained.  The prices of housing and utilities rose by 2.6%Y in September (versus 2.1%Y in August). The government announced a rise in electricity tariffs in August, and these were already reflected in the September CPI numbers. The prices of recreation and culture rose by 2.1%Y in September (versus 1.6%Y in August), as Koreans sought more recreational activities during the Chuseok holidays. 

Miscellaneous goods/service prices also increased by 5.1%Y in September (versus 4.6%Y in August). Clothing prices, however, saw lower inflation in the month, rising by 2.3%Y (versus 2.7%Y in August). The likely reason could be more promotional sales due to the change of season and during the Chuseok holidays.  Transport and education prices also reported lower inflation at 1.8%Y (versus 2.8%Y in August) and 2.1%Y (versus 2.2%Y in August), respectively.

Rate Hike Warranted to Anchor Inflation Expectations

After the surprise jump in September CPI, we think it is time for the BoK to raise the policy interest rate at a faster pace. The BoK raised the policy seven-day repo rate by 25bp to 2.25% in July this year, which marked the start of the rate normalisation cycle. The market and we expected another rate hike in September, but the BoK held the rate unchanged for the month. The next meeting is scheduled for October 14, and we think that a rate hike is very likely. We believe that Korea's strong economic fundamentals have already justified a faster rate normalisation pace. Given that food prices are most closely linked with people's everyday life, the price hikes could very easily give rise to higher inflation expectations among consumers, especially in a lax monetary environment with abundant liquidity. We think that rate hikes are needed to anchor inflation expectations, and we expect a rate hike of 25bp at the October meeting.