diff --git a/nano/core_test/active_transactions.cpp b/nano/core_test/active_transactions.cpp
index fc3e1d97b5..f09117c6e8 100644
--- a/nano/core_test/active_transactions.cpp
+++ b/nano/core_test/active_transactions.cpp
@@ -1458,3 +1458,73 @@ TEST (active_transactions, vacancy)
 	ASSERT_EQ (1, node.active.vacancy ());
 	ASSERT_EQ (0, node.active.size ());
 }
+
+// Ensure transactions in excess of capacity are removed in fifo order
+TEST (active_transactions, fifo)
+{
+	nano::system system;
+	nano::node_config config{ nano::get_available_port (), system.logging };
+	config.active_elections_size = 1;
+	auto & node = *system.add_node (config);
+	nano::keypair key0;
+	nano::keypair key1;
+	nano::state_block_builder builder;
+	// Construct two pending entries that can be received simultaneously
+	auto send0 = builder.make_block ()
+				 .account (nano::dev_genesis_key.pub)
+				 .previous (nano::genesis_hash)
+				 .representative (nano::dev_genesis_key.pub)
+				 .link (key0.pub)
+				 .balance (nano::genesis_amount - 1)
+				 .sign (nano::dev_genesis_key.prv, nano::dev_genesis_key.pub)
+				 .work (*system.work.generate (nano::genesis_hash))
+				 .build_shared ();
+	ASSERT_EQ (nano::process_result::progress, node.process (*send0).code);
+	nano::blocks_confirm (node, { send0 }, true);
+	ASSERT_TIMELY (1s, node.block_confirmed (send0->hash ()));
+	ASSERT_TIMELY (1s, node.active.empty ());
+	auto send1 = builder.make_block ()
+				 .account (nano::dev_genesis_key.pub)
+				 .previous (send0->hash ())
+				 .representative (nano::dev_genesis_key.pub)
+				 .link (key1.pub)
+				 .balance (nano::genesis_amount - 2)
+				 .sign (nano::dev_genesis_key.prv, nano::dev_genesis_key.pub)
+				 .work (*system.work.generate (send0->hash ()))
+				 .build_shared ();
+	ASSERT_EQ (nano::process_result::progress, node.process (*send1).code);
+	nano::blocks_confirm (node, { send1 }, true);
+	ASSERT_TIMELY (1s, node.block_confirmed (send1->hash ()));
+	ASSERT_TIMELY (1s, node.active.empty ());
+
+	auto receive0 = builder.make_block ()
+					.account (key0.pub)
+					.previous (0)
+					.representative (nano::dev_genesis_key.pub)
+					.link (send0->hash ())
+					.balance (1)
+					.sign (key0.prv, key0.pub)
+					.work (*system.work.generate (key0.pub))
+					.build_shared ();
+	ASSERT_EQ (nano::process_result::progress, node.process (*receive0).code);
+	auto receive1 = builder.make_block ()
+					.account (key1.pub)
+					.previous (0)
+					.representative (nano::dev_genesis_key.pub)
+					.link (send1->hash ())
+					.balance (1)
+					.sign (key1.prv, key1.pub)
+					.work (*system.work.generate (key1.pub))
+					.build_shared ();
+	ASSERT_EQ (nano::process_result::progress, node.process (*receive1).code);
+	node.scheduler.manual (receive0);
+	// Ensure first transaction becomes active
+	ASSERT_TIMELY (1s, node.active.election (receive0->qualified_root ()) != nullptr);
+	node.scheduler.manual (receive1);
+	// Ensure second transaction becomes active
+	ASSERT_TIMELY (1s, node.active.election (receive1->qualified_root ()) != nullptr);
+	// Ensure excess transactions get trimmed
+	ASSERT_TIMELY (1s, node.active.size () == 1);
+	// Ensure the surviving transaction is the least recently inserted
+	ASSERT_TIMELY (1s, node.active.election (receive1->qualified_root ()) != nullptr);
+}
diff --git a/nano/core_test/election_scheduler.cpp b/nano/core_test/election_scheduler.cpp
index 98ba24993b..c95dc39292 100644
--- a/nano/core_test/election_scheduler.cpp
+++ b/nano/core_test/election_scheduler.cpp
@@ -142,7 +142,8 @@ TEST (election_scheduler, flush_vacancy)
 				.sign (nano::dev_genesis_key.prv, nano::dev_genesis_key.pub)
 				.work (*system.work.generate (nano::genesis_hash))
 				.build_shared ();
-	node.scheduler.manual (send);
+	ASSERT_EQ (nano::process_result::progress, node.process (*send).code);
+	node.scheduler.activate (nano::dev_genesis_key.pub, node.store.tx_begin_read ());
 	// Ensure this call does not block, even though no elections can be activated.
 	node.scheduler.flush ();
 	ASSERT_EQ (0, node.active.size ());
diff --git a/nano/node/active_transactions.cpp b/nano/node/active_transactions.cpp
index fe70504450..39d95dcb7c 100644
--- a/nano/node/active_transactions.cpp
+++ b/nano/node/active_transactions.cpp
@@ -317,9 +317,7 @@ void nano::active_transactions::request_confirm (nano::unique_lock<nano::mutex>
 	{
 		bool const confirmed_l (election_l->confirmed ());
 
-		unconfirmed_count_l += !confirmed_l;
-		bool const overflow_l (unconfirmed_count_l > node.config.active_elections_size && election_l->election_start < election_ttl_cutoff_l);
-		if (overflow_l || election_l->transition_time (solicitor))
+		if (election_l->transition_time (solicitor))
 		{
 			if (election_l->optimistic () && election_l->failed ())
 			{
@@ -1046,6 +1044,16 @@ void nano::active_transactions::erase_hash (nano::block_hash const & hash_a)
 	debug_assert (erased == 1);
 }
 
+void nano::active_transactions::erase_oldest ()
+{
+	nano::unique_lock<nano::mutex> lock (mutex);
+	if (!roots.empty ())
+	{
+		auto item = roots.get<tag_random_access> ().front ();
+		cleanup_election (lock, *item.election);
+	}
+}
+
 bool nano::active_transactions::empty ()
 {
 	nano::lock_guard<nano::mutex> lock (mutex);
diff --git a/nano/node/active_transactions.hpp b/nano/node/active_transactions.hpp
index 32337d0ee2..24e8940ac8 100644
--- a/nano/node/active_transactions.hpp
+++ b/nano/node/active_transactions.hpp
@@ -167,6 +167,7 @@ class active_transactions final
 	std::vector<std::shared_ptr<nano::election>> list_active (size_t = std::numeric_limits<size_t>::max ());
 	void erase (nano::block const &);
 	void erase_hash (nano::block_hash const &);
+	void erase_oldest ();
 	bool empty ();
 	size_t size ();
 	void stop ();
diff --git a/nano/node/election_scheduler.cpp b/nano/node/election_scheduler.cpp
index 71a9991ad7..b7fc5f1fac 100644
--- a/nano/node/election_scheduler.cpp
+++ b/nano/node/election_scheduler.cpp
@@ -87,6 +87,21 @@ size_t nano::election_scheduler::priority_queue_size () const
 	return priority.size ();
 }
 
+bool nano::election_scheduler::priority_queue_predicate () const
+{
+	return node.active.vacancy () > 0 && !priority.empty ();
+}
+
+bool nano::election_scheduler::manual_queue_predicate () const
+{
+	return !manual_queue.empty ();
+}
+
+bool nano::election_scheduler::overfill_predicate () const
+{
+	return node.active.vacancy () < 0;
+}
+
 void nano::election_scheduler::run ()
 {
 	nano::thread_role::set (nano::thread_role::name::election_scheduler);
@@ -94,15 +109,23 @@ void nano::election_scheduler::run ()
 	while (!stopped)
 	{
 		condition.wait (lock, [this] () {
-			auto vacancy = node.active.vacancy ();
-			auto has_vacancy = vacancy > 0;
-			auto available = !priority.empty () || !manual_queue.empty ();
-			return stopped || (has_vacancy && available);
+			return stopped || priority_queue_predicate () || manual_queue_predicate () || overfill_predicate ();
 		});
 		debug_assert ((std::this_thread::yield (), true)); // Introduce some random delay in debug builds
 		if (!stopped)
 		{
-			if (!priority.empty ())
+			if (overfill_predicate ())
+			{
+				node.active.erase_oldest ();
+			}
+			else if (manual_queue_predicate ())
+			{
+				auto const [block, previous_balance, election_behavior, confirmation_action] = manual_queue.front ();
+				nano::unique_lock<nano::mutex> lock2 (node.active.mutex);
+				node.active.insert_impl (lock2, block, previous_balance, election_behavior, confirmation_action);
+				manual_queue.pop_front ();
+			}
+			else if (priority_queue_predicate ())
 			{
 				auto block = priority.top ();
 				std::shared_ptr<nano::election> election;
@@ -113,15 +136,6 @@ void nano::election_scheduler::run ()
 					election->transition_active ();
 				}
 				priority.pop ();
-				++priority_queued;
-			}
-			if (!manual_queue.empty ())
-			{
-				auto const [block, previous_balance, election_behavior, confirmation_action] = manual_queue.front ();
-				nano::unique_lock<nano::mutex> lock2 (node.active.mutex);
-				node.active.insert_impl (lock2, block, previous_balance, election_behavior, confirmation_action);
-				manual_queue.pop_front ();
-				++manual_queued;
 			}
 			notify ();
 		}
diff --git a/nano/node/election_scheduler.hpp b/nano/node/election_scheduler.hpp
index 8baf7a7e10..1a2889f2d1 100644
--- a/nano/node/election_scheduler.hpp
+++ b/nano/node/election_scheduler.hpp
@@ -36,10 +36,11 @@ class election_scheduler final
 private:
 	void run ();
 	bool empty_locked () const;
+	bool priority_queue_predicate () const;
+	bool manual_queue_predicate () const;
+	bool overfill_predicate () const;
 	nano::prioritization priority;
-	uint64_t priority_queued{ 0 };
 	std::deque<std::tuple<std::shared_ptr<nano::block>, boost::optional<nano::uint128_t>, nano::election_behavior, std::function<void (std::shared_ptr<nano::block>)>>> manual_queue;
-	uint64_t manual_queued{ 0 };
 	nano::node & node;
 	bool stopped;
 	nano::condition_variable condition;