diff --git a/src/backend/optimizer/path/allpaths.c b/src/backend/optimizer/path/allpaths.c
index aa78c0af0cd..057b4b79ebb 100644
--- a/src/backend/optimizer/path/allpaths.c
+++ b/src/backend/optimizer/path/allpaths.c
@@ -3079,8 +3079,7 @@ generate_gather_paths(PlannerInfo *root, RelOptInfo *rel, bool override_rows)
 	 * of partial_pathlist because of the way add_partial_path works.
 	 */
 	cheapest_partial_path = linitial(rel->partial_pathlist);
-	rows =
-		cheapest_partial_path->rows * cheapest_partial_path->parallel_workers;
+	rows = compute_gather_rows(cheapest_partial_path);
 	simple_gather_path = (Path *)
 		create_gather_path(root, rel, cheapest_partial_path, rel->reltarget,
 						   NULL, rowsp);
@@ -3098,7 +3097,7 @@ generate_gather_paths(PlannerInfo *root, RelOptInfo *rel, bool override_rows)
 		if (subpath->pathkeys == NIL)
 			continue;
 
-		rows = subpath->rows * subpath->parallel_workers;
+		rows = compute_gather_rows(subpath);
 		path = create_gather_merge_path(root, rel, subpath, rel->reltarget,
 										subpath->pathkeys, NULL, rowsp);
 		add_path(rel, &path->path);
@@ -3282,7 +3281,6 @@ generate_useful_gather_paths(PlannerInfo *root, RelOptInfo *rel, bool override_r
 													subpath,
 													useful_pathkeys,
 													-1.0);
-				rows = subpath->rows * subpath->parallel_workers;
 			}
 			else
 				subpath = (Path *) create_incremental_sort_path(root,
@@ -3291,6 +3289,7 @@ generate_useful_gather_paths(PlannerInfo *root, RelOptInfo *rel, bool override_r
 																useful_pathkeys,
 																presorted_keys,
 																-1);
+			rows = compute_gather_rows(subpath);
 			path = create_gather_merge_path(root, rel,
 											subpath,
 											rel->reltarget,
diff --git a/src/backend/optimizer/path/costsize.c b/src/backend/optimizer/path/costsize.c
index 2021c481b46..79991b19807 100644
--- a/src/backend/optimizer/path/costsize.c
+++ b/src/backend/optimizer/path/costsize.c
@@ -6473,3 +6473,21 @@ compute_bitmap_pages(PlannerInfo *root, RelOptInfo *baserel,
 
 	return pages_fetched;
 }
+
+/*
+ * compute_gather_rows
+ *	  Estimate number of rows for gather (merge) nodes.
+ *
+ * In a parallel plan, each worker's row estimate is determined by dividing the
+ * total number of rows by parallel_divisor, which accounts for the leader's
+ * contribution in addition to the number of workers.  Accordingly, when
+ * estimating the number of rows for gather (merge) nodes, we multiply the rows
+ * per worker by the same parallel_divisor to undo the division.
+ */
+double
+compute_gather_rows(Path *path)
+{
+	Assert(path->parallel_workers > 0);
+
+	return clamp_row_est(path->rows * get_parallel_divisor(path));
+}
diff --git a/src/backend/optimizer/plan/planner.c b/src/backend/optimizer/plan/planner.c
index 4711f912390..948afd90948 100644
--- a/src/backend/optimizer/plan/planner.c
+++ b/src/backend/optimizer/plan/planner.c
@@ -5370,8 +5370,7 @@ create_ordered_paths(PlannerInfo *root,
 																	root->sort_pathkeys,
 																	presorted_keys,
 																	limit_tuples);
-			total_groups = input_path->rows *
-				input_path->parallel_workers;
+			total_groups = compute_gather_rows(sorted_path);
 			sorted_path = (Path *)
 				create_gather_merge_path(root, ordered_rel,
 										 sorted_path,
@@ -7543,8 +7542,6 @@ gather_grouping_paths(PlannerInfo *root, RelOptInfo *rel)
 			(presorted_keys == 0 || !enable_incremental_sort))
 			continue;
 
-		total_groups = path->rows * path->parallel_workers;
-
 		/*
 		 * We've no need to consider both a sort and incremental sort. We'll
 		 * just do a sort if there are no presorted keys and an incremental
@@ -7561,7 +7558,7 @@ gather_grouping_paths(PlannerInfo *root, RelOptInfo *rel)
 														 groupby_pathkeys,
 														 presorted_keys,
 														 -1.0);
-
+		total_groups = compute_gather_rows(path);
 		path = (Path *)
 			create_gather_merge_path(root,
 									 rel,
diff --git a/src/backend/optimizer/util/pathnode.c b/src/backend/optimizer/util/pathnode.c
index c42742d2c7b..d1c4e1a6aa7 100644
--- a/src/backend/optimizer/util/pathnode.c
+++ b/src/backend/optimizer/util/pathnode.c
@@ -1899,7 +1899,6 @@ create_gather_merge_path(PlannerInfo *root, RelOptInfo *rel, Path *subpath,
 	pathnode->num_workers = subpath->parallel_workers;
 	pathnode->path.pathkeys = pathkeys;
 	pathnode->path.pathtarget = target ? target : rel->reltarget;
-	pathnode->path.rows += subpath->rows;
 
 	if (pathkeys_contained_in(pathkeys, subpath->pathkeys))
 	{
diff --git a/src/include/optimizer/cost.h b/src/include/optimizer/cost.h
index b1c51a4e70f..57861bfb446 100644
--- a/src/include/optimizer/cost.h
+++ b/src/include/optimizer/cost.h
@@ -212,5 +212,6 @@ extern PathTarget *set_pathtarget_cost_width(PlannerInfo *root, PathTarget *targ
 extern double compute_bitmap_pages(PlannerInfo *root, RelOptInfo *baserel,
 								   Path *bitmapqual, double loop_count,
 								   Cost *cost_p, double *tuples_p);
+extern double compute_gather_rows(Path *path);
 
 #endif							/* COST_H */
diff --git a/src/test/regress/expected/join_hash.out b/src/test/regress/expected/join_hash.out
index 262fa71ed8d..4fc34a0e72a 100644
--- a/src/test/regress/expected/join_hash.out
+++ b/src/test/regress/expected/join_hash.out
@@ -508,18 +508,17 @@ set local hash_mem_multiplier = 1.0;
 set local enable_parallel_hash = on;
 explain (costs off)
   select count(*) from simple r join extremely_skewed s using (id);
-                              QUERY PLAN                               
------------------------------------------------------------------------
- Finalize Aggregate
+                           QUERY PLAN                            
+-----------------------------------------------------------------
+ Aggregate
    ->  Gather
          Workers Planned: 1
-         ->  Partial Aggregate
-               ->  Parallel Hash Join
-                     Hash Cond: (r.id = s.id)
-                     ->  Parallel Seq Scan on simple r
-                     ->  Parallel Hash
-                           ->  Parallel Seq Scan on extremely_skewed s
-(9 rows)
+         ->  Parallel Hash Join
+               Hash Cond: (r.id = s.id)
+               ->  Parallel Seq Scan on simple r
+               ->  Parallel Hash
+                     ->  Parallel Seq Scan on extremely_skewed s
+(8 rows)
 
 select count(*) from simple r join extremely_skewed s using (id);
  count