Why Timing Matters
A policy rate changes on a particular date, but the economic conditions that prompted the change may have begun earlier, and the change may affect spending and inflation only later. Meanwhile, official data are released after the periods they describe and can be revised. A decision based only on the latest fully observed inflation and output data can therefore respond to an economy that has already moved on.
Monetary policy timing is often described using three lags: recognition, decision, and transmission. They mark different intervals, not interchangeable explanations for a single delay. Separating them helps a policy maker identify what is known, when an action can take effect, and when its broader effects might become visible. The actual lengths are uncertain; a lag schedule is a planning assumption, not a promise.
From Economic Development to Policy Effects
The stages can be laid out as a timeline. A development may begin before it appears clearly in published data. Recognition may require several observations, and the decision may occur at a scheduled meeting rather than as soon as a signal appears. After the decision, the operating change takes effect; only then does the transmission lag begin. A complete timeline should label each date or period rather than using “the lag” without specifying which interval it means.
A change in demand, costs, or other conditions starts to influence the economy, but may not yet be measurable in available data.
Data releases and other indicators provide enough evidence to assess whether the change is persistent or meaningful. Initial estimates may later be revised.
Policy makers choose a response, and the resulting policy change takes effect. The decision and effective dates need not coincide.
Economic responses emerge over time. Effects on activity and inflation need not occur together or have the same timing.
Forecast-based policy decisions use information about expected future conditions as well as observed data. This is important because waiting until every current indicator is complete may mean acting after the underlying development has already changed. But forecasts are uncertain: an early response based on a mistaken interpretation can also be poorly timed. The practical question is not simply whether to act early or late; it is how to weigh incomplete evidence against the consequences of delay.
Worked Example: Separating the Stages on a Calendar
Suppose a hypothetical weakening in demand begins on February 10, in Quarter 1. A central bank considers the change sufficiently clear by June 15, in Quarter 2, decides on August 5, in Quarter 3, and makes the policy change effective on August 8, also in Quarter 3. Assume broader activity effects begin three quarters after the effective date. Identify the stages and the quarter in which those effects begin.
Method. Use the stated dates for recognition, decision, and implementation, then count transmission quarters from the effective quarter. The February development date is not the start of the transmission clock.
Recognition lag. From February 10 to June 15 is 4 months and 5 days. This is the interval from the beginning of the development to sufficient recognition in this scenario.
Decision lag. From June 15 to August 5 is 51 days. The decision occurs after recognition rather than on the date the development began.
Implementation lag. From August 5 to August 8 is 3 days. The policy change takes effect on August 8.
Transmission timing. Define horizon 0 as the quarter containing the effective date. August 8 is in Quarter 3, so Quarter 3 is horizon 0. Quarter 4 is horizon 1, Quarter 1 of the following year is horizon 2, and Quarter 2 of the following year is horizon 3. The assumed broader effects begin in Quarter 2 of the following year.
Check. The intervals are distinct: recognition runs from February to June, decision from June to August, implementation from August 5 to August 8, and the assumed three-quarter transmission horizon starts at the effective quarter. Counting three quarters forward from horizon 0 gives the stated answer.
Reading a Lag Schedule
A lag schedule represents an assumed pattern of effects after a policy change takes effect. For a simple illustrative model, let \(\Delta r\) be a change in the policy rate measured in percentage points. Let \(\Delta y_h\) be the associated change in an output measure, and \(\Delta \pi_h\) the associated change in the inflation rate, both also measured in percentage points, at horizon \(h\). Define \(b_h\) and \(c_h\) as nonnegative response magnitudes for a tightening. The assumed responses are:
The minus signs encode the illustrative direction: a positive rate increase is associated with downward pressure on output and inflation in this schedule. The magnitudes \(b_h\) and \(c_h\) are not universal constants. Their values, timing, and even the observed pattern can differ across circumstances. A response can be small early and larger later, and inflation may respond on a different schedule from output. A schedule is therefore a way to reason about timing, not a claim that the economy follows an exact mechanical path.
Worked Example: Mapping a Rate Increase to Quarterly Responses
A hypothetical policy rate rises by 0.50 percentage points effective April 1, in Quarter 2. An illustrative schedule gives the following nonnegative magnitudes: at horizons 0 through 3, \(b_h\) equals 0.10, 0.30, 0.60, and 0.80, while \(c_h\) equals 0.05, 0.10, 0.20, and 0.35. Find the scheduled output and inflation changes at each horizon.
Method. Set horizon 0 to the quarter containing the effective date, multiply each magnitude by the rate change, and apply the minus sign in the response equations. The output and inflation entries are changes in percentage points, not percentage changes in the magnitudes themselves.
At horizon 0, Quarter 2, the calculations are \(\Delta y_0=-0.10(0.50)=-0.05\) and \(\Delta \pi_0=-0.05(0.50)=-0.025\) percentage points.
At horizon 1, Quarter 3, they are \(\Delta y_1=-0.30(0.50)=-0.15\) and \(\Delta \pi_1=-0.10(0.50)=-0.05\) percentage points.
At horizon 2, Quarter 4, they are \(\Delta y_2=-0.60(0.50)=-0.30\) and \(\Delta \pi_2=-0.20(0.50)=-0.10\) percentage points.
At horizon 3, Quarter 1 of the following year, they are \(\Delta y_3=-0.80(0.50)=-0.40\) and \(\Delta \pi_3=-0.35(0.50)=-0.175\) percentage points.
Check. The largest scheduled magnitudes are at horizon 3, so the largest downward responses in this schedule are also there. The rate change is positive and the response magnitudes are nonnegative, so each calculated response is negative as specified.
Why Delayed Responses Accumulate
When policy changes occur in more than one quarter, the response observed in a given quarter may reflect several earlier actions. A useful justification for adding these effects is to suppose, for illustration, that each policy change has its own schedule and that the model is locally linear: the response to two changes together equals the sum of their separate responses. If the change in quarter \(s\) is \(\Delta r_s\), its output contribution \(h\) quarters later is \(-b_h\Delta r_s\). In quarter \(t\), the elapsed horizon for that action is \(t-s\). Adding the contributions gives:
This follows by applying the single-change schedule separately to each policy move and summing the resulting effects in the quarter being considered. It assumes a common schedule and linear addition; it does not account for interactions, changing economic conditions, or uncertainty in the schedule. Those limits matter: the formula organizes a scenario, but is not a guarantee of the observed outcome.
Worked Example: Accounting for Two Policy Changes
Suppose the policy rate is initially 3.0%. It increases by 0.40 percentage points in Quarter 1, to 3.40%, and by a further 0.20 percentage points in Quarter 2, to 3.60%. These hypothetical decisions are made while some current data remain incomplete. Assume the same illustrative schedule applies to both changes: \(b_0=0.10\), \(b_1=0.30\), \(b_2=0.50\), and \(c_h=b_h/2\). Find the combined scheduled changes in output and inflation in Quarters 1, 2, and 3.
Method. Treat each increase as a separate policy change and add its contribution in each quarter. The second increase is 0.20 percentage points, not the new total rate of 3.60%.
In Quarter 1, only the first increase has taken effect, at horizon 0. Output changes by \(-0.10(0.40)=-0.04\) percentage points. Since \(c_0=0.05\), inflation changes by \(-0.05(0.40)=-0.02\) percentage points.
In Quarter 2, the first increase is at horizon 1 and the second is at horizon 0. Output changes by \(-0.30(0.40)-0.10(0.20)=-0.12-0.02=-0.14\) percentage points. Inflation changes by \(-0.15(0.40)-0.05(0.20)=-0.06-0.01=-0.07\) percentage points.
In Quarter 3, the first increase is at horizon 2 and the second is at horizon 1. Output changes by \(-0.50(0.40)-0.30(0.20)=-0.20-0.06=-0.26\) percentage points. Inflation changes by \(-0.25(0.40)-0.15(0.20)=-0.10-0.03=-0.13\) percentage points.
Check. The Quarter 3 result includes both the delayed effect of the Quarter 1 action and the earlier effect of the Quarter 2 action. Omitting either contribution would understate the combined scheduled response under the stated assumptions.
Common Timing and Sign Errors
A frequent timing error is to count the effective quarter as horizon 1. If a change takes effect in Quarter 2 and a response is assumed at horizon 2, Quarter 2 is horizon 0, Quarter 3 is horizon 1, and Quarter 4 is horizon 2. Calling Quarter 3 the horizon-2 quarter counts the effective quarter inconsistently.
A separate sign error occurs when a response table already provides negative effects and the minus sign in the formula is applied again. For example, with \(\Delta r=0.50\) and \(b_1=0.30\), the correct calculation is \(\Delta y_1=-0.30(0.50)=-0.15\) percentage points. The incorrect calculation \(-(-0.30)(0.50)=+0.15\) treats the nonnegative magnitude \(b_1\) as though it were already a signed response. Keep magnitudes nonnegative and use the sign in the equation.
Finally, do not confuse the date the economic change began with the date policy became effective. Recognition and decision lags occur before effectiveness; transmission horizons are counted from effectiveness. Forecasts can help policy makers account for developments not yet fully shown in current data, but forecasts and lag schedules should be treated as uncertain inputs rather than exact predictions.
Check Your Understanding
For each scenario, identify the relevant timing or calculate the scheduled response. Treat stated quarter horizons as beginning with horizon 0 in the quarter the policy change takes effect.
- A hypothetical rate change takes effect in Quarter 4. A specified response begins three quarters after effectiveness. In which quarter and at what horizon does it begin? (Answer: Quarter 3 of the following year, at horizon 3.)
- A development begins on April 10 in Quarter 2, is recognized on July 1 in Quarter 3, and prompts a decision on August 15. The change takes effect on September 1, still in Quarter 3. If broader effects begin at horizon 2, when do they begin? (Answer: Quarter 1 of the following year, at horizon 2.)
- A 0.60-percentage-point rate increase has \(b_2=0.45\) and \(c_2=0.20\). Find the scheduled output and inflation changes at horizon 2. (Answer: \(\Delta y_2=-0.27\) and \(\Delta\pi_2=-0.12\) percentage points.)
- The rate rises by 0.50 percentage points in Quarter 2 and by a further 0.25 percentage points in Quarter 3. If \(b_0=0.20\) and \(b_1=0.40\), what is the combined scheduled output change in Quarter 3? (Answer: \(-0.40(0.50)-0.20(0.25)=-0.25\) percentage points.)
- A policy change becomes effective on October 1, in Quarter 4, and an assumed response begins at horizon 3. State the response quarter. (Answer: Quarter 3 of the following year, at horizon 3.)